What is the output limitation of HHL-type algorithms?

Question: In the quantum linear systems setting Ax=b, what is a central caveat of HHL-style algorithms? A) They output all entries of x explicitly in classical memory in polylogarithmic time B) They typically output a quantum state proportional to x, so extracting the full classical vector may destroy the claimed speedup C) They require A to be dense and ill-conditioned to obtain advantage D) They avoid dependence on the condition number by using phase estimation Correct: B Explanation: HHL-style algorithms can prepare a quantum state proportional to the solution vector under assumptions such as sparsity and good conditioning. But reading out all vector entries classically can require many measurements, so the speedup is meaningful mainly when the solution is consumed quantumly or only certain observables are needed. Topic: advanced quantum computing / HHL / quantum linear systems